Pressure driven membrane separation processes are known wherein organic molecules or inorganic ionic solutes in aqueous solutions are concentrated or separated to various degrees by the application of a positive osmotic pressure to one side of a filtration membrane. Examples of such pressures are reverse osmosis (RO), ultrafiltration (UF) and nanofiltration (NF). These pressure driven membrane processes employ a cross-flow mode of operation wherein only a portion of a feed solution (F) is collected as a permeate solution (P) and the rest is collected as a pass solution (C). In this specification and claims, the exit process stream from the nanofiltration module, which stream has not passed through the membrane is referred to as the "pass stream". This stream is often referred to by practitioners in the membrane filtration art as the "concentrate" stream.
In the case of separation of two solutes A and B, say, NaCl and Na.sub.2 SO.sub.4, the efficiency of the separation process is identified by the following parameters: ##EQU1## wherein A!.sub.F is solute A concentration in feed solution;
A!.sub.P is solute A concentration in permeate solution; PA1 F.sub.P is permeate solution flow; and PA1 F.sub.F is feed solution flow. PA1 (a) Desal-5 Membrane Product Application Note, publication of Desalination Systems, Inc. (Escondido, Calif.), April 1991, wherein the Figure on page E-19.3 shows NaCl rejection in the 55 to 85% range; PA1 (b) NT70 Membrane, Product Specification, publication of Filmtec Corp. (Minneapolis, Minn.), cites Rejection of 60%; and PA1 (c) "Membrane Handbook", ed. By W. S. W. Ho and K. K. Sirkar, Van Norstrand Reinhold, N.Y. 1992 at Table 23.2. "Characteristics of Selected Nanofiltration Membranes", cites NaCl % Rejection of: 80% for NF50 membrane (Filmtec), 45% for NF40 membrane (Filmtec), 50% for NTF-7250 membrane (Nitto), 47% for Desal-5 membrane (Desalination Systems), and 55% for SU200HF membrane (Toray). PA1 Removal of multivalent metals from brine. Also from acids such as H.sub.2 SO.sub.4, HNO.sub.3, HCl, HF or mixtures thereof such as galvanic wastewater, metal cleaning, metal etching and the like. PA1 Separation of NaCl from Na.sub.2 SO.sub.4 and Na.sub.2 CO.sub.3 in a dissolved precipitator catch from a recovery boiler in a Kraft pulp mill. Removal of chlorides is required to reduce the corrosivity of recovered process chemical streams within a Kraft mill which is substantially closed, i.e. effluent free. PA1 Purification of fertilizer grade of orthophosphoric acid from heavy metals to make it suitable for technical application, i.e. upgrading to technical grade acid. PA1 Recovery of H.sub.2 SO.sub.4 and HNO.sub.3 from spent nitration acid. Here the nitrated organic byproducts remain in the pass liquor stream, while purer and desired acids are collected as a permeate. PA1 Separation of phenolic salts from a product rinse water during production of nitrobenzenes, nitrotoluenes, nitroxylenes and other nitroorganic compounds. PA1 Segregation of sodium sesquisulfate solution, Na.sub.3 H(SO.sub.4).sub.2 into Na.sub.2 SO.sub.4 in the pass liquor stream and NaHSO.sub.4 in the permeate. The latter could be used within a pulp mill as an acid, e.g. for generation of ClO.sub.2 or in an acidulation step to produce a tall oil. PA1 Fractionation of White Liquor into a Na.sub.2 S-rich pass liquor and a NaOH-rich permeate fraction.
When a separation of solute A from solute B is required, a high % Rejection of solute A and a low Rejection of solute B, or vice versa, high % Recovery and high Permeate Flux is desired.
Nanofiltration membranes are structurally very similar to reverse osmosis membranes in that chemically, they, typically, are crosslinked aromatic polyamides, which are cast as a thin "skin layer", on top of a microporous polymer sheet support to form a composite membrane structure. The separation properties of the membrane are controlled by the pore size and electrical charge of the "skin layer". Such a membrane structure is usually referred to as a thin film composite (TFC). However, unlike RO membranes, the NF membranes are characterized in having a larger pore size in its "skin layer" and a net negative electrical charge inside the individual pores. This negative charge is responsible for rejection of anionic species, according to the anion surface charge density. Accordingly, divalent anions, such as SO.sub.4.sup.=, are more strongly rejected than monovalent ones, such as Cl.sup.-. Commercial NF membranes are available from known suppliers of RO and other pressure drive membranes. Examples include: Desal-5 membrane (Desalination Systems, Escondido, Calif.), NF70, NF50, NF40 and NF40HF membranes (FilmTec Corp., Minneapolis, Minn.), SU 600 membrane (Toray, Japan), and NRT 7450 and NTR 7250 membranes (Nitto Electric, Japan). The NF membranes are, typically, packaged as membrane modules. A so-called "spiral wound" module is most popular, but other membrane module configurations, such as tubular membranes enclosed in a shell or plate- and-frame type, are also known.
Nanofiltration is characterized by a fractionation capacity for organic solutes with a molecular "cut-off" range of about 300 g/mol; and a fractionation capacity for multivalent vs. monovalent ions, which is especially pronounced for anions.
Nanofiltration membranes have been reported to show no or little rejection of low molecular weight organic molecules, such as, methanol, ethanol and ethyleneglycol, but a significant rejection of higher molecular weight organic species, such as glucose. Among inorganic ionic solutes, low to medium rejection has been reported for simple 1:1 electrolytes, such as NaCl or NaNO.sub.3 and high rejection of other electrolytes where multivalent ionic species are involved, such as Na.sub.2 SO.sub.4, MgCl.sub.2 and FeCl.sub.3. Such a characteristic differentiates NF from RO which rejects all ionic species, and from ultrafiltration (UF), which does not reject ionic species and only rejects organic compounds with molecular weights, typically, in excess of 1,000 g/mol.
Sodium chloride (C.sup.-) finite % Rejections have been published in the following publications, namely:
During the NF process, a minimum pressure equal to the osmotic pressure difference between the feed/pass liquor on one side and the permeate liquor on the other side of the membrane must be applied since osmotic pressure is a function of the ionic strengths of the two streams. In the case of separation of a multivalent solute, such as Na.sub.2 SO.sub.4, from a monovalent one, such as NaCl, the osmotic pressure difference is moderated by the low NaCl rejection. Usually, a pressure in excess of the osmotic pressure difference is employed to achieve practical permeate flux. In view of lower NaCl rejection, NF has been used successfully for removal of sulfate and the hardness cations, Ca.sup.2+ and Mg.sup.2+ from brackish waters and even seawater, without the necessity to excessively pressurize the feed stream. The reported typical pressure range for NF is 80 to 300 psi, although membrane elements are designed to withstand pressures of up to 1,000 psi.
Reported uses of NF include the aforesaid water softening, removal of dissolved multivalent ions such as Ra.sup.2+, reduction of silica as a part of feedwater conditioning for a subsequent RO step or removal of medium of medium molecular weight organic compounds. It has also been demonstrated that high rejection of ionic species could be obtained by proper conditioning of the stream, i.e. by changing its pH. Thus, effective removal (rejection) of carbonate anion could be achieved by adjusting the pH of the feed solution to about 12, to ensure that carbonate would predominantly exist as CO.sub.3.sup.=, which anion is more strongly rejected by the NF membrane than the HCO.sub.3.sup.- anion.
Dissolved or suspended silica in brine feed for chloralkali processes, especially the so-called membrane chloralkali process, presents a problem in that the silica forms scale on the surface or in the interior of the ion exchange membrane separator. This causes the cell voltage and, hence, power consumption to increase. In general, in the membrane chloralkali process, the concentration of silica in the feed brine should not exceed 10 ppm, although even a lower level may be needed if some other contaminants, such as Al.sup.3+, are present, since these contaminants enhance the scaling capacity of silica.
In other types of chloralkali and in sodium chlorate manufacturing processes, silica, if present in the feed brine, also leads to insoluble deposits on the anode which also leads to increased cell voltage and a premature wear of the anode coating. In general, however, in these processes, somewhat higher levels of silica, e.g. 30 ppm or more can still be tolerated.
Silica is recognized as a difficult contaminant to remove from water and/or brine. In chloralkali practice, it is usually removed by the addition of MgCl.sub.2 or FeCl.sub.3 to brine, followed by pH adjustment to precipitate the respective metal hydroxide in a form of a floc. This freshly formed floc is an effective absorber for dissolved silica, which may then be separated from brine by e.g. filtration. One method combining aeration of brine, to convert Fe(II) present therein to Fe(III), which then forms Fe(OH).sub.3 floc is described in U.S. Pat. No. 4,405,463.
Use of strongly basic anion exchange membranes for silica removal from feedwater has been reported. However, the literature also recognizes that, in that case since there is a substantial background of other salts, the selectivity of the IX resin towards silica is greatly reduced.
Product literature from FilmTec Corp., Minneapolis, Minn., describes the removal of silica from feedwater with a NF70 nanofiltration membrane, as part of a pretreatment for a subsequent RO step. A reduction of silica concentration in feedwater from 400 ppm to 50-60 ppm has been mentioned. The literature is silent, however, on the use of NF methods for silica removal from higher concentration salt solutions, such as chloralkali brine.
Sodium chlorate is generally prepared by the electrolysis of sodium chloride to produce chlorine, sodium hydroxide and hydrogen. The chlorine and sodium hydroxide are immediately reacted to form sodium hypochlorite, which is then converted to chlorate and chloride under controlled conditions of pH and temperature.
In a related chemical process, chlorine and caustic soda are prepared in an electrolytic cell, which contains a membrane to prevent chlorine and caustic soda reacting and the separated chemicals are removed.
The sodium chloride salt used to prepare the brine for electrolysis to sodium chlorate generally contains impurities which, depending on the nature of the impurity and production techniques employed, can give rise to plant operational problems familiar to those skilled in the art. The means of controlling these impurities are varied and include, purging them out of the system into alternative processes or to the drain, precipitation by conversion to insoluble salts, crystallization or ion exchange treatment. The control of anionic impurities presents more complex problems than that of cationic impurities.
Sulfate ion is a common ingredient in commercial salt. When such salt is used directly, or in the form of a brine solution, and specific steps are not taken to remove the sulfate, the sulfate enters the electrolytic system. Sulfate ion maintains its identity under the conditions in the electrolytic system and, thus, accumulates and progressively increases in concentration in the system unless removed in some manner. In chlorate plants producing a liquor product, the sulfate ion will leave with the product liquor. In plants producing only crystalline chlorate, the sulfate remains in the mother liquor after the crystallization of the chlorate, and is recycled to the cells. Over time, the concentration of sulfate ion will increase and adversely affect electrolysis and cause operational problems due to localized precipitation in the electrolytic cells. Within the chloralkali circuit, the sodium sulfate will concentrate and adversely effect the membrane, which divides the anolyte (brine) from the catholyte (caustic soda).
It is industrially desirable that sodium sulfate levels in concentrated brine, e.g., 300 g/l NaCl, be reduced to at least 20 g/l in chlorate production and 10 g/l in chloralkali production.
U.S. Pat. No. 4,702,805, Burkell and Warren, issued Oct. 27, 1987, describes an improved method for the control of sulfate in an electrolyte stream in a crystalline chlorate plant, whereby the sulfate is crystallized out. In the production of crystalline sodium chlorate according to U.S. Pat. No. 4,702,805, sodium chlorate is crystallized from a sodium chlorate rich liquor. The crystals are removed to provide a mother liquor comprising principally sodium chlorate and sodium chloride, together with other components, including sulfate and dichromate ions. A portion of the mother liquor is cooled to a temperature to effect crystallization of a portion of the sulfate as sodium sulfate in admixture with sodium chlorate. The crystallized admixture is removed and the resulting spent method liquor is recycled to the electrolytic process.
It has been found subsequently, that the crystallized admixture of sulfate and chlorate obtained from typical commercial liquors according to the process of U.S. Pat. No. 4,702,805, may be discoloured yellow owing to the unexpected occlusion of a chromium component in the crystals. The discolouration cannot be removed by washing the separated admixture with liquors in which the crystallized sulfate and chlorate are insoluble. It will be appreciated that the presence of chromium in such a sulfate product is detrimental in subsequent utilization of this product and, thus, this represents a limitation to the process as taught in U.S. Pat. No. 4,702,805.
U.S. Pat. No. 4,636,376--Maloney and Carbaugh, issued Jan. 13, 1987, discloses removing sulfate from aqueous chromate-containing sodium chlorate liquor without simultaneous removal of significant quantities of chromate. The chromate and sulfate-containing chlorate liquor having a pH in the range of about 2.0 to about 6.0 is treated with a calcium-containing material at a temperature of between about 40.degree. C. and 95.degree. C., for between 2 and 24 hours to form a sulfate-containing precipitate. The precipitate is predominantly glauberite, Na.sub.2 Ca(SO.sub.4).sub.2. However, the addition of calcium cations requires the additional expense and effort of the treatment and removal of all excess calcium ions. It is known that calcium ions may form an unwanted deposit on the cathodes which increases the electrical resistance of the cells and adds to operating costs. It is, typically, necessary to remove calcium ions by means of ion exchange resins.
U.S. Pat. No. 5,093,089--Alford and Mok, issued Mar. 3, 1992, describes an improved version of the selective crystallization process of aforesaid U.S. Pat. No. 4,702,805, wherein process conditions are selected to provide precipitation of sulfate substantially free of chromium contaminant.
Typically, organic anion exchange resins have a low selectivity for sulfate anions in the presence of a large excess of chlorine ions. U.S. Pat. No. 4,415,677 describes a sulfate ion absorption method, but which method has disadvantages. The method consists of removing sulfate ions from brine by a macroporous ion exchange resin composite having polymeric zirconium hydrous oxide contained in a vessel. This method is not economical because the efficiency is low and a large amount of expensive cation exchange resin is required for carrying polymeric zirconium hydrous oxide. Further, the polymeric zirconium hydrous oxide adsorbing sulfate ions comes into contact with acidic brine containing sulfate ions, resulting in loss of polymeric zirconium hydrous oxide due to acid-induced dissolution. Soluble zirconyl ions precipitates as hydroxide in the lower portion of the vessel to clog flow paths.
U.S. Pat. No. 4,556,463--Minz and Vajna, issued Dec. 3, 1984, describes a process to decrease sulfate concentration levels in brine solutions using an organic ion exchange material with brine streams under carefully controlled dilutive conditions.
U.S. Pat. No. 5,071,563--Shiga et al., issued Dec. 10, 1991, describes the selective adsorption of sulfate anions from brine solutions using zirconium hydrous oxide slurry under acidic conditions. The ion exchange compound may be regenerated by treatment with alkali.
Japanese Patent Kokai No. 04321514-A, published Nov. 11, 1992 to Kaneka Corporation describes the selective adsorption of sulfate anions from brine solutions using cerium hydroxide slurry under acidic conditions. The ion exchange compound may be regenerated by treatment with alkali.
Japanese Patent Kokai No. 04338110-A--Kaneka Corporation, published Nov. 25, 1992, describes the selective adsorption of sulfate anions from brine solutions using titanium hydrous oxide slurry under acidic conditions. The ion exchange compound may be regenerated by treatment with alkali.
Japanese Patent Kokai No. 04334553-A--Kaneka, published Nov. 11, 1992, describes the removal of sulfate ions from brine using ion-adsorbing cakes in a slurry.
There still remains, however, a need for an improved, cost-effective, practical method for the removal of sulfate, silica and chromium (VI) ions from alkali metal halide solutions, particularly, from sodium chloride solutions used in the electrolytic production of sodium chlorate and chlorine/caustic soda.